February 2007

New Center Paves Way for Supercomputer Research and Development

By Zak M. Salih

"Research in the field of high-performance computing at the University received a powerful boost with the recent establishment at GW of the National Science Foundation Center for High-Performance Reconfigurable Computing (CHREC). Funded by the National Science Foundation (NSF), its Industry/University Cooperative Research Centers program (I/UCRC), and more than 20 government and industrial organizations, CHREC will be one of the world’s most advanced research centers for high- performance reconfigurable computers, known as supercomputers.

In addition to GW, the University of Florida will host a CHREC location. Two other universities— Virginia Tech and Brigham Young—also will become partner institutions pending approval of their proposals by the NSF.

The field of high-performance reconfigurable computing involves the use of computers with both a standard processor and pieces of hardware that can be reconfigured to execute more specific tasks. The last decade has seen advancement in the study and use of these types of machines.

The center’s goals include establishing the nation’s first multidisciplinary research center in reconfigurable high-performance computing; cost-effectively supporting the research needs of other partners; enhancing the educational experience for graduate and undergraduate students; and advancing the technology and knowledge in the field with an emphasis on commercial relevance.

A key component of the new center is its cooperation with numerous research partners and founding members in government and industry, including the National Security Agency, Intel, Oak Ridge National Lab, Arctic Region Supercomputing Center, Intel Coporation, Silicon Graphics, Linux Networks, Hewlett-Packard, and the National Cancer Institute. These members, instrumental in helping CHREC secure initial funding from the NSF, also provide significant cash resources through annual membership fees in CHREC in addition to state-of-the-art equipment. In return, they play a hands-on role in the research process and have full access to the center’s work.

According to Tarek El-Ghazawi, GW professor of engineering and applied science and co-director of CHREC, the center’s corporate partners are very enthusiastic about its work. “Not too many of these big corporations have yet started their own research and development efforts in the field,” he says.

Each of the four partner institutions will focus on a particular area: GW on high-performance computing, the University of Florida on embedded systems, Virginia Tech on user interface, and Brigham Young on debugging and other chip-level issues. “It’s a very complimentary group,” says El-Ghazawi.

The GW site currently has three projects planned for 2007-08. El-Ghazawi
and a team of professors

and graduate students will explore and streamline computer applications to make them run as fast as possible, focus on biomedical imaging and gene sequencing in a study of portable applications used in hospitals, and examine how to productively program high-performance reconfigurable computers.

“GW and its partner institution are doing breakthrough research that is industrially relevant and will help the United States maintain competitiveness within the computer industry,” says Alex Schwartzkopf, director of the I/UCRC program at the NSF.

Alan George, professor of electrical and computer engineering at the University of Florida and director of CHREC, agrees that GW’s participation is integral to the center’s work. “Because of its academic and research excellence in high-performance and reconfigurable computing, GW is widely respected and serves a vital role in the success of this new NSF center,” he says.

El-Ghazawi hopes CHREC will create momentum to attract more top faculty and students and strengthen the University’s expertise in the field. Already, other universities including Washington University, Rice University, and the University of California, Riverside, have expressed interest in the center.

“We definitely would like to see it become self- sustained once the initial NSF support ends,” El-Ghazawi says. “Also, from a structural and operational perspective, we’d like to use this as a vehicle to advance our own education and research agenda at GW.”

Send feedback to: bygeorge@gwu.edu